[rfc-i] Thoughts on the Independent Stream
sm at resistor.net
Sat Oct 22 11:54:39 PDT 2011
At 07:47 22-10-2011, Eric Burger wrote:
>A.k.a., tilting at windmills.
>Consider this as comments to draft-iab-ise-model-01.
>I would offer that the Independent Stream is obsolete. The need to
>create a venue to publish non-IETF standards publications is totally
>irrelevant now that we have this thing called the World Wide
>Web. There is a lot of confusion in the marketplace that an RFC is
>an IETF publication. MANY people who publish Independent
>Submissions do so with the explicit purpose of leveraging that
>confusion. I would offer the Independent Stream is destabilizing,
>evil, and something not to be supported.
There are documents such as RFC 4357 which are not published within
the IETF Stream. Even from an IETF perspective, it is good to have
such specifications documented in an archival series. The thing
called the World Wide Wide is not really an archival series.
>The ONLY work product of the IETF is an RFC. Independent
>Submissions dilute the RFC brand. It is no solution for us to go
>out and say, "Yes, you think the work product of the IETF is an RFC,
>but it might not really be an IETF work product." That makes the
>RFC brand even more diminished than it already is with the presence
>of the Independent Stream.
This argument could be distilled as the RFC brand belongs to the IETF
and a document can only be published as a RFC if the IETF agrees to
it. Olaf mentioned the IAB and IRTF streams. We could also argue
that these two streams dilute the RFC brand.
>I would like to point out that the average cost of RFC publication
>is $1,200 per RFC. However, the paid staff involvement in an
>Independent Stream RFC, per RFC 5620, is much higher than that of a
>Standards Track RFC. For the latter, the review is performed by
>volunteers. For Independent Stream documents, per RFC 5620 and as
>outlined in draft-iab-ise-model-01 in Section 2.2 bullet 2 (hence
>the relevance to this discussion), the ISE does the review,
>shepherding, and coordination. Instead of being done by volunteers,
>we pay the ISE. Therefore, I would expect the actual cost of
>publication of an Independent Stream RFC would be considerably
>greater than $1,200.
Is the ISE being paid? Based on previous discussions, I thought that
it was a volunteer position.
>I am not a total naysayer. I understand the value of having an
>archival publication spot for non-IETF work that we want to let IETF
>participants and the Internet community at large to have
>access. Here are some proposals for how to achieve that goal.
>If the purpose of taking independent documents is to truly gather
>requests for comment, then I would offer the existing I-D mechanism
>works perfectly. This is more especially so given that while I-D's
>expire in 6 months in theory, they actually are archived indefinitely.
Wouldn't that also be applicable for Individual Submissions?
>If the purpose of taking independent documents is to document
>hard-to-get or possibly ephemeral, proprietary protocol
>specifications, we could use the IANA function or some other
>mechanism, such as piggy-backing on the RFC repository
>infrastructure, to stand up an archival web server. This will cost
>considerably less than $1,200 per document. So long as we do not
>call them RFCs, it means the only thing we need is to get copyright
>permission to put the document on-line. There would be no confusion
>with a given non-IETF document with an IETF RFC. Note that this
>eradicates the need for an Independent Submissions Editor, as there
>is nothing to edit. This eradicates Responsibilities (Section 2.2)
>1, 2, 3, 4, and 6. Responsibility 5 just goes away, as there can be
>no errata to a non-IETF publication. If there is an error in the
>base document, the owner of that document can always submit a new
>version. Responsibility 7 does not require a human, and as such does
>not need to be done by the ISE.
There is an interesting point in what you mentioned above. If the
cost is $1,200 per RFC, would it be better to stick to original RFC +
errata instead of publishing a new version?
>So, if all Responsibilities go away, there is nothing for an ISE to
>do. If there is nothing for an ISE to do, there is no need to
>document what an ISE does. Since there is no need to document what
>an ISE does, there is no need to publish draft-iab-ise-model-01 as an RFC.
And you save $1,200. :-)
>If we are really tied to having an Independent Stream that produces
>RFCs, then I would offer that we follow the IEEE model. That model
>says that we will publish a document for free if it is less than,
>say, 4 pages. We will be happy to publish longer documents, but the
>person requesting publication would need to pay a per page fee. For
>comparison, the IEEE/ACM Transactions on Networking charges $220 per
>page over 10 pages. J-SAC charges $220 per page over 7 pages. If
>we are going to support confusion, there is no reason not to profit
>from it and at least generate some revenue to support the positive
>work that we do.
I expect that this revenue stream will be shared with people who
review the document. :-)
As we are discussing about RFCs, I have a question. The RFC author
gets the glory. What do the people sending in comments get?
If I had to summarize your message, I would put it as "Is there a
justification for the costs involved in producing RFCs". Should
there even be RFCs given that I-Ds are archived indefinitely? Using
I-Ds only would reduce the costs. BTW, if the RFC Editor did not
have anything to edit (author is responsible for doing editorial
fixes), that would reduce the costs.
RFC services cost around $886,700. If 1% of that is for Independent
Stream RFCs and savings have to be made, you don't chase that 1%
unless it's about government expenditure where pennies cause outrage
as people can be distracted from the big picture. Although I rate
what you said as constructive, it would be better to focus on the 99%
as it might translate into larger savings.
More information about the rfc-interest